I've been getting up the curve on Using the STM32 cortex-m3's full-speed library (Datasheet, FS Library documentation, Jan's book, beyondlogic, etc) and have read a LOT of great posts and I think I have a pretty good handle on my implementation.
In my application, I will have a lot of reports (big report descriptor). It looks like the default implementation reserves 168 bytes for the Control endpoint transmission buffer in the packet memory.
If my descriptor is larger than 168 bytes, the library doesn't automatically handle splitting the transfers correct? I'm thinking I might need to modify the core functions to write in chunks of 64?
Thanks! Thomas